JPH0566540B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to recognition of floc formation status in a water treatment plant by measuring the particle size and distribution of floc in a floc formation pond (mixing pond) of a water treatment plant using image processing technology. Regarding the method.

[Background of the invention]

At water treatment plants, the suspended particles in the raw water are small in size, so the process is to agglomerate them into flocs, and then let the flocs settle. For this reason, it is essential to monitor the flocs in the floc formation pond (mixing pond).

Traditionally, flocs have been visually monitored several times a day by water treatment plant maintenance managers. Because it relies on visual inspection, the criteria for judgment are subjective and qualitative;
The drawback is that the monitoring results are not easily reflected in driving operations. Furthermore, since the monitoring frequency is discontinuous, countermeasures against agglomeration failures are delayed, which increases troubles.

In contrast, recently, a method has been adopted in which an industrial television camera (ITV) is used to monitor the flocs in the floc formation pond. However, even in this case, monitoring relies on human vision;
It remains subjective and discontinuous.

In order to improve these shortcomings, JP-A-54-
As described in Japanese Patent No. 143296, a method has also been proposed in which a photoelectric conversion device is used to extract an electrical signal according to the shape of a flock. However, in practical use, there are the following problems.

In order to recognize the floc formation status online, it is necessary to install a photoelectric conversion device with a watertight structure in the water of the floc formation pond at the water treatment plant.

At this time, the flocs floating in the floc formation pond accumulate on the photoelectric conversion device having a watertight structure, and if they accumulate to a certain extent, a phenomenon occurs in which the flocs slide down from the watertight mechanism due to their own weight or the influence of the water flow in the floc formation pond. At this time, if a cluster of accumulated flocs enters and passes through the floc observation space by the photoelectric conversion device, the image processing device will recognize that a huge floc has been formed, making it impossible to correctly recognize the floc formation situation. Also,
Generally, flock formation ponds are constructed in a state open to the atmosphere, so foreign objects such as dead leaves and trash thrown into the air by meteorological phenomena such as strong winds land on the flock formation pond. Even if this foreign matter settles in the floc formation pond and passes through the floc observation space by the photoelectric conversion device, the image processing device will recognize that a huge floc has been formed, making it impossible to recognize the correct floc formation situation. It disappears. The recognition information on the floc formation status is applied to control of the amount of flocculant injected into the rapid mixing pond, control of the rotation speed of the flocculator in the flocculation pond, and the like. In this case, for example, the amount of flocculant injected should be increased because the flocs that are actually formed are small, but it is mistakenly recognized that huge flocs have been formed, so it is mistakenly judged that the amount of flocculant injection should be decreased. As a result, flocs cannot be formed properly and become a burden on the filtration ponds, etc. in subsequent processes.

[Purpose of the invention]

An object of the present invention is to provide a method for recognizing floc formation status in a water purification plant that can accurately and continuously measure the particle size and distribution of flocs in a floc formation pond (mixing pond).

[Summary of the invention]

A feature of the present invention is that when the area of individual flocs in a floc formation pond is measured by image processing and the floc formation status is determined by determining, for example, the particle size distribution, the area of one floc is The reason is that the floc is not recognized as a floc when it is recognized as a predetermined value or more.

[Embodiments of the invention]

FIG. 1 shows an embodiment of the present invention.

In Figure 1, the taken raw water is #10
It is guided to the rapid mixing basin 20 via the. A stirrer 21 is provided in the rapid mixing pond 20. The agitator 21 stirs the flocculant injected from the flocculant injector 22 . Flotsk formation pond 30
Stirring paddles 31A, 31B, and 31C are provided. The stirring paddles are rotationally driven by stirring motors 32A, 32B, and 32C, respectively. Stirring paddles 31A, 31B, 31C
A rectifying wall 33A with many holes bored between the
It is separated by 33B. The fine flocs flowing from the rapid mixing basin 20 are sequentially stirred in the floc forming basin 30 by stirring paddles 31A, 31B, and 31C. As a result, the micro flocs collide and coalesce to become flocs 34. While the flocs 34 pass through the floc formation pond 30, their particle size gradually increases.

The imaging device 100 converts the image of the flock 34 into an electrical signal using an industrial television camera (not shown). The image processing device 110 performs image processing on the flock 34 based on the electrical signal obtained from the imaging device 100, and calculates the area S of the flock 34.

Flock 34 formed in the floc formation pond 30
flows into the settling basin 40 and is precipitated and removed. The supernatant water from which the flocs 34 have been removed is injected into the filter basin 50. In the filtration tank 50, remaining minute flocs that were not removed in the settling tank 40 are filtered out. The supernatant water flowing out from the filtration basin 50 is drained into a drainage basin (not shown).
Water is supplied to consumers via a reservoir (not shown), etc.

FIG. 2 shows an imaging device 100 and an image processing device 110.
The detailed configuration is shown below.

An industrial television camera (ITV) 130 fixed in an airtight container 120 uses a close-up lens 131 to
The flocs 34 in the floc formation pond 30 can be seen through the observation window 121 made of a transparent material such as glass.
Enlarge and recognize the image. Wiper drive device 123
The wiper 122 driven by the observation window 12
It operates periodically to remove dirt from the surface of 1.
A back screen 124 is provided to accurately recognize flock groups with high contrast. The screen 124 is installed in front of the observation window 121 via back screen fixtures 124A and 124B fixed to the airtight container 120. The back screen 124 is preferably a dark color in order to accurately recognize white flocks with high contrast. The light-shielding cover 142 is provided to darken the surroundings and provide illuminance under a certain condition only by the illumination device 140. By providing the light-shielding cover 142, the image of the flock group is not affected by changes in surrounding illumination. It is more practical to install a plurality of illumination devices 140 to irradiate the flock from multiple directions. Note that the lighting device 140 can also be controlled by the lighting controller 141 according to changes in ambient illuminance without providing the light-shielding cover 142.
Jiyama boards 125A and 125B are floc formation pond 3
This is provided in order to suppress the movement of water as much as possible within the interior and to avoid blocking the illumination. By suppressing the movement of water, images of flocks in a flowing state can be recognized with high accuracy.

Flock image information captured by ITV 130 is sent to image recognition device 150 via ITV controller 132. Image recognition control device 16
0 controls the frequency of flock recognition by the image recognition device 150. Generally, the condition of floc formation does not change rapidly in a short period of time. Therefore, it is sufficient to perform the flock monitoring operation once every 10 minutes to once every hour.

The process by which the image information of the flock group is subjected to signal processing within the image recognition device 150 will be described below.

FIG. 3a shows the image plane of the flocks recognized by the ITV 130. Since the flock group is white, the brightness level is high, and the background back screen 1
24 is black, so the brightness level of the water is low. Since the flock group is a grayscale image, the boundary between the flock 34 and the water portion is not clear in reality. FIG. 3a only shows the outline of the flock group for the sake of simplicity.

FIG. 3b shows the luminance level distribution scanned along line AA' on the screen shown in FIG. 3a. The brightness level is
The brightness level is displayed in 256 levels, with the brightness level increasing toward the top of the vertical axis and decreasing toward the bottom. Since the flock 34 is white, the luminance level increases. In FIG. 3b, the downwardly trough portion represents the floc. The luminance distribution shown in FIG. 3b is binarized based on the luminance level threshold of line BB'. A pixel whose luminance level is higher than the threshold value is set to a "1" level, and a pixel whose luminance level is below the threshold value is set to a "0" level. FIG. 3c shows a signal waveform obtained by binarizing the screen shown in FIG. 3a along line AA'.

By subjecting the entire screen of FIG. 3a to the aforementioned binarization process, a binarized screen as shown in FIG. 3d is obtained. 300 in FIG. 3d indicates each pixel. Each flock has 30 pixels
The area of each floc ( _{S i} , i=1, 2, . . . n) is calculated from the total sum of 0 (N s i , i=1 . . . n) using the following formula.

S _i =C ₁ ×N _Si (1) C ₁ : Actual area to which one pixel corresponds A signal of the calculated area is output from the image recognition device 150.

Through the process described above, the area S _i of each floc
is measured.

Next, the process of floc formation will be explained.

At water treatment plants, the suspended particles in raw water are small in size, so the process is to agglomerate them to form flocs, and then settle the flocs.

In the rapid mixing pond 20, when a flocculant, which is a positively charged colloid, is injected into raw water suspended fine particles, which are negatively charged colloids, the suspended fine particles are combined with the flocculant. are tied together to form a microflock. Next, flock formation pond 3
At 0, the micro flocs flow and collide with each other due to the rotation of the stirring paddles 31A to 31C of the flocculator to form flocs. The flocs in the floc formation pond 30 are agitated by the flocculator, thereby increasing the frequency of floc-to-flock collisions. Collisions promote growth into larger flocs. On the other hand, for flocs that have grown to a certain size, the agitation by the flocculator acts as a type of shearing force, resulting in the destruction of large flocs. For this reason, empirically, the particle size of the flocs formed in the floc formation pond is approximately several mm even if they grow. This is well known by those skilled in the art. A predetermined area S _e that is larger than the area that can be considered as a large floc is set in the arithmetic unit 210, and the area S _i of the floc obtained by the image processing device 110 and the predetermined set area S _c are calculated by the arithmetic unit 210. Compare with. The arithmetic unit 210 recognizes that the object is a floc when S _i <S _c , and recognizes that it is not a floc but a foreign object when S _i >S _e .

Through the processing described above, it becomes possible to remove foreign objects and recognize only flocs.

A specific example will be described below using FIGS. 4 to 9.

Note that FIGS. 4, 6, and 8 are diagrams showing only the parts necessary for explaining the imaging device 100 in FIG. 2.

FIG. 4 shows the jamb board 125A of the imaging device 100,
A mass of flocs 500 accumulated on the board 125B is caused by its own weight or the influence of water flow, etc.
125A and 125B, and enters and passes through the flock observation space 400. FIG. At this time, image information as shown in FIG. 5a is obtained.
This image information is binarized using a constant luminance level threshold value and is shown in FIG. 5b. Each flock in FIG. 5b is numbered No. 1, 2, . . . 5. next
The area (S ₁ to _{S 5} ) of each of No. 1 to 5 flocs is calculated and compared with a predetermined set area S _e . In the case of FIG. 5b, assuming that only the area _S3 of the No. 3 floc is recognized as larger than the predetermined set area S _e , the No. 3 object is recognized as not being a floc.

FIG. 6 shows an example of a foreign object 501 landing on or entering the floc formation pond 30, and FIG. 7b shows the image information in FIG. 7a binarized and numbered. In the case of Figure 7b, No.4
The object is recognized as not a flock.

FIG. 8 shows an example in which the imaging device 100 is provided with a wiper 122 to remove dirt from the surface of the observation window 121. Wiper 122 operation and ITV 130
The image information capture operations are taken into consideration so that they do not overlap. However, when image information existing on the wiper 122 screen is captured due to a malfunction of the wiper 122, the image information is
It will look like figure a. FIG. 9b shows the image information in FIG. 9a that has been binarized and numbered. In the case of FIG. 9b, object No. 2 is recognized as not being a flock.

FIG. 10 shows an example of a processing flow in accordance with the above description.

The ITV 130 in FIG. 2 captures the image information of the flock 34 and sends it to the image recognition device 150 via the ITV controller 132 (step S1).
The image information of the flock sent to the image recognition device 150 in step S2 is binarized by the image recognition device 150, numbered (step S3), and then the area S _i is calculated in step S4. step
In S5, the arithmetic unit 210 compares the area S _i of each flock sent from the image recognition device 150 with the set area S _e set in the arithmetic unit 210. When S _i <S _e , the object is recognized as a floc (step S6), and the floc formation status is grasped by calculating, for example, the particle size distribution (step S7). When S _i >S _e , the object is recognized as a foreign object rather than a floc, and is removed without being included in the information to be processed (step S8), and an alarm etc. is output (step S9).

By repeating these steps S1 to S9, the flock observation space 40 of the imaging device 100 is
It becomes possible to accurately and continuously recognize the state of floc formation without being affected by foreign matter that has entered or passed through.

In addition, by using the particle size distribution obtained in step S7, etc., it is possible to control the amount of coagulant injected, the amount of alkali agent injected, etc., as well as control the rotational speed of the stirring paddles 31A to 31C.

Next, as shown in FIG. 11a, a case will be described in which a portion of the foreign object 501 that has landed on or entered the floc formation pond 30 is obtained as image information in the floc observation space 400.

A part of the flock 34 and the foreign object 501 are located in the flock observation space 400 as shown at 502 in FIG. 11a.
The 11th image is obtained by binarizing the image information obtained as image information using a certain brightness level threshold.
Figure b. Number each object in Figure 11b as No. 1, 2, 3, and 4. Figure 11b
The area S ₂ of object No. 2 inside is the object 50 in Figure 11a.
Although it is part of 1, the predetermined set area
If it is smaller than S _e , it will be mistakenly recognized as a flock. Similarly, the area S ₄ of object No. 4 in Figure 11b
Although it is a part of the flock 502 in FIG. 11a, it is mistakenly recognized as a flock of area S ₄ .

As described above, depending on the timing of taking in the image information, a part of the foreign object or a part of the floc may be obtained as image information in the peripheral area of the floc observation space, so that erroneous recognition may occur.

This problem can be solved as follows.

Before numbering the binarized image information b obtained by binarizing the image information in FIG. Both write a "1" level (same level as the flock) with a width of one pixel or more. By doing this, all the objects 501 and 502, part of which was obtained as image information, are connected to the peripheral part of the flock observation space and recognized as one object, as shown in FIG. 11c. Ru. Therefore, when numbering is done for Figure 11c, all objects 501 and 502 whose parts are obtained as image information in the peripheral area of the flock observation space are recognized as one object No. 1. be done. By recognizing this object No. 1 as not being a floc, objects other than objects whose entire image has been obtained as image information in the floc observation space are excluded from the objective of understanding the floc formation status. This makes it possible to accurately recognize flocks.

〔Effect of the invention〕

As explained above, according to the present invention, it is possible to accurately recognize only flocs without being affected by foreign objects that have entered or passed through the floc observation space of the imaging device. , objectively, quantitatively and with high precision.
Online automatic measurement is possible. This makes it possible to save labor and improve reliability in water treatment plant maintenance and management, and even to apply it to control systems.

[Brief explanation of drawings]

FIG. 1 is a configuration diagram showing one embodiment of the present invention, and FIG.
The figure is a detailed configuration diagram of the main part of Figure 1, Figure 3 is an explanatory diagram of image signal processing, Figures 4 to 9 are explanatory diagrams of specific examples of the present invention, Figure 10 is an operation flow diagram, and Figure 11 is an illustration of the image signal processing. It is an explanatory diagram of image recognition. 10...Water landing well, 20...Rapid mixing pond, 30...
Float formation pond, 100... Imaging device, 110...
...Image processing device, 210...Arithmetic device.

Claims (1)

[Claims] 1. In a water treatment plant where a flocculant is injected into raw water and then introduced into a floc-forming pond to grow flocs, the area of each individual floc in the floc-forming pond is measured by image processing, and the area of the flocs is determined from the area of the measured flocs. 1. A method for recognizing floc formation status in a water purification plant, characterized in that when recognizing the floc formation status, if the individual floc area exceeds a predetermined value, the flocs are not recognized as flocs.